diff options
| -rw-r--r-- | test/test_cache.py | 2 | ||||
| -rw-r--r-- | test/test_jsinterp.py | 63 | ||||
| -rw-r--r-- | test/test_youtube_signature.py | 64 | ||||
| -rwxr-xr-x | youtube_dl/YoutubeDL.py | 37 | ||||
| -rw-r--r-- | youtube_dl/cache.py | 69 | ||||
| -rw-r--r-- | youtube_dl/compat.py | 85 | ||||
| -rw-r--r-- | youtube_dl/extractor/common.py | 90 | ||||
| -rw-r--r-- | youtube_dl/extractor/youtube.py | 115 | ||||
| -rw-r--r-- | youtube_dl/jsinterp.py | 278 | ||||
| -rw-r--r-- | youtube_dl/utils.py | 8 | ||||
| -rw-r--r-- | youtube_dl/version.py | 2 |
11 files changed, 555 insertions, 258 deletions
diff --git a/test/test_cache.py b/test/test_cache.py index 931074aa1..21dfb6cfc 100644 --- a/test/test_cache.py +++ b/test/test_cache.py @@ -63,7 +63,7 @@ class TestCache(unittest.TestCase): obj = {'x': 1, 'y': ['รค', '\\a', True]} c.store('test_cache', 'k.', obj) self.assertEqual(c.load('test_cache', 'k.', min_ver='1970.01.01'), obj) - new_version = '.'.join(('%d' % ((v + 1) if i == 0 else v, )) for i, v in enumerate(version_tuple(__version__))) + new_version = '.'.join(('%0.2d' % ((v + 1) if i == 0 else v, )) for i, v in enumerate(version_tuple(__version__))) self.assertIs(c.load('test_cache', 'k.', min_ver=new_version), None) diff --git a/test/test_jsinterp.py b/test/test_jsinterp.py index 3c9650ab6..479cb43a0 100644 --- a/test/test_jsinterp.py +++ b/test/test_jsinterp.py @@ -7,6 +7,7 @@ from __future__ import unicode_literals import os import sys import unittest + sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__)))) import math @@ -146,6 +147,25 @@ class TestJSInterpreter(unittest.TestCase): # https://github.com/ytdl-org/youtube-dl/issues/32815 self._test('function f(){return 0 - 7 * - 6;}', 42) + def test_bitwise_operators_typecast(self): + # madness + self._test('function f(){return null << 5}', 0) + self._test('function f(){return undefined >> 5}', 0) + self._test('function f(){return 42 << NaN}', 42) + self._test('function f(){return 42 << Infinity}', 42) + self._test('function f(){return 0.0 << null}', 0) + self._test('function f(){return NaN << 42}', 0) + self._test('function f(){return "21.9" << 1}', 42) + self._test('function f(){return true << "5";}', 32) + self._test('function f(){return true << true;}', 2) + self._test('function f(){return "19" & "21.9";}', 17) + self._test('function f(){return "19" & false;}', 0) + self._test('function f(){return "11.0" >> "2.1";}', 2) + self._test('function f(){return 5 ^ 9;}', 12) + self._test('function f(){return 0.0 << NaN}', 0) + self._test('function f(){return null << undefined}', 0) + self._test('function f(){return 21 << 4294967297}', 42) + def test_array_access(self): self._test('function f(){var x = [1,2,3]; x[0] = 4; x[0] = 5; x[2.0] = 7; return x;}', [5, 2, 7]) @@ -160,6 +180,7 @@ class TestJSInterpreter(unittest.TestCase): self._test('function f(){var x = 20; x = 30 + 1; return x;}', 31) self._test('function f(){var x = 20; x += 30 + 1; return x;}', 51) self._test('function f(){var x = 20; x -= 30 + 1; return x;}', -11) + self._test('function f(){var x = 2; var y = ["a", "b"]; y[x%y["length"]]="z"; return y}', ['z', 'b']) def test_comments(self): self._test(''' @@ -351,6 +372,13 @@ class TestJSInterpreter(unittest.TestCase): self._test('function f() { a=5; return (a -= 1, a+=3, a); }', 7) self._test('function f() { return (l=[0,1,2,3], function(a, b){return a+b})((l[1], l[2]), l[3]) }', 5) + def test_not(self): + self._test('function f() { return ! undefined; }', True) + self._test('function f() { return !0; }', True) + self._test('function f() { return !!0; }', False) + self._test('function f() { return ![]; }', False) + self._test('function f() { return !0 !== false; }', True) + def test_void(self): self._test('function f() { return void 42; }', JS_Undefined) @@ -435,6 +463,7 @@ class TestJSInterpreter(unittest.TestCase): def test_regex(self): self._test('function f() { let a=/,,[/,913,/](,)}/; }', None) + self._test('function f() { let a=/,,[/,913,/](,)}/; return a.source; }', ',,[/,913,/](,)}') jsi = JSInterpreter(''' function x() { let a=/,,[/,913,/](,)}/; "".replace(a, ""); return a; } @@ -482,25 +511,6 @@ class TestJSInterpreter(unittest.TestCase): self._test('function f(){return -524999584 << 5}', 379882496) self._test('function f(){return 1236566549 << 5}', 915423904) - def test_bitwise_operators_typecast(self): - # madness - self._test('function f(){return null << 5}', 0) - self._test('function f(){return undefined >> 5}', 0) - self._test('function f(){return 42 << NaN}', 42) - self._test('function f(){return 42 << Infinity}', 42) - self._test('function f(){return 0.0 << null}', 0) - self._test('function f(){return NaN << 42}', 0) - self._test('function f(){return "21.9" << 1}', 42) - self._test('function f(){return 21 << 4294967297}', 42) - self._test('function f(){return true << "5";}', 32) - self._test('function f(){return true << true;}', 2) - self._test('function f(){return "19" & "21.9";}', 17) - self._test('function f(){return "19" & false;}', 0) - self._test('function f(){return "11.0" >> "2.1";}', 2) - self._test('function f(){return 5 ^ 9;}', 12) - self._test('function f(){return 0.0 << NaN}', 0) - self._test('function f(){return null << undefined}', 0) - def test_negative(self): self._test('function f(){return 2 * -2.0 ;}', -4) self._test('function f(){return 2 - - -2 ;}', 0) @@ -543,6 +553,8 @@ class TestJSInterpreter(unittest.TestCase): test_result = list('test') tests = [ 'function f(a, b){return a.split(b)}', + 'function f(a, b){return a["split"](b)}', + 'function f(a, b){let x = ["split"]; return a[x[0]](b)}', 'function f(a, b){return String.prototype.split.call(a, b)}', 'function f(a, b){return String.prototype.split.apply(a, [b])}', ] @@ -593,6 +605,9 @@ class TestJSInterpreter(unittest.TestCase): self._test('function f(){return "012345678".slice(-1, 1)}', '') self._test('function f(){return "012345678".slice(-3, -1)}', '67') + def test_splice(self): + self._test('function f(){var T = ["0", "1", "2"]; T["splice"](2, 1, "0")[0]; return T }', ['0', '1', '0']) + def test_pop(self): # pop self._test('function f(){var a = [0, 1, 2, 3, 4, 5, 6, 7, 8]; return [a.pop(), a]}', @@ -627,6 +642,16 @@ class TestJSInterpreter(unittest.TestCase): 'return [ret.length, ret[0][0], ret[1][1], ret[0][2]]}', [2, 4, 1, [4, 2]]) + def test_extract_function(self): + jsi = JSInterpreter('function a(b) { return b + 1; }') + func = jsi.extract_function('a') + self.assertEqual(func([2]), 3) + + def test_extract_function_with_global_stack(self): + jsi = JSInterpreter('function c(d) { return d + e + f + g; }') + func = jsi.extract_function('c', {'e': 10}, {'f': 100, 'g': 1000}) + self.assertEqual(func([1]), 1111) + if __name__ == '__main__': unittest.main() diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py index 8fdcce8d4..a2d3a41ae 100644 --- a/test/test_youtube_signature.py +++ b/test/test_youtube_signature.py @@ -95,10 +95,50 @@ _SIG_TESTS = [ '0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpz2ICs6EVdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', ), ( + 'https://www.youtube.com/s/player/363db69b/player_ias_tce.vflset/en_US/base.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + '0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpz2ICs6EVdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + ), + ( 'https://www.youtube.com/s/player/4fcd6e4a/player_ias.vflset/en_US/base.js', '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', 'wAOAOq0QJ8ARAIgXmPlOPSBkkUs1bYFYlJCfe29xx8q7v1pDL0QwbdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJoOySqa0', ), + ( + 'https://www.youtube.com/s/player/4fcd6e4a/player_ias_tce.vflset/en_US/base.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + 'wAOAOq0QJ8ARAIgXmPlOPSBkkUs1bYFYlJCfe29xx8q7v1pDL0QwbdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJoOySqa0', + ), + ( + 'https://www.youtube.com/s/player/20830619/player_ias.vflset/en_US/base.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + '7AOq0QJ8wRAIgXmPlOPSBkkAs1bYFYlJCfe29xx8jOv1pDL0Q2bdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJoOySqa0qaw', + ), + ( + 'https://www.youtube.com/s/player/20830619/player_ias_tce.vflset/en_US/base.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + '7AOq0QJ8wRAIgXmPlOPSBkkAs1bYFYlJCfe29xx8jOv1pDL0Q2bdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJoOySqa0qaw', + ), + ( + 'https://www.youtube.com/s/player/20830619/player-plasma-ias-phone-en_US.vflset/base.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + '7AOq0QJ8wRAIgXmPlOPSBkkAs1bYFYlJCfe29xx8jOv1pDL0Q2bdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJoOySqa0qaw', + ), + ( + 'https://www.youtube.com/s/player/20830619/player-plasma-ias-tablet-en_US.vflset/base.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + '7AOq0QJ8wRAIgXmPlOPSBkkAs1bYFYlJCfe29xx8jOv1pDL0Q2bdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJoOySqa0qaw', + ), + ( + 'https://www.youtube.com/s/player/8a8ac953/player_ias_tce.vflset/en_US/base.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + 'IAOAOq0QJ8wRAAgXmPlOPSBkkUs1bYFYlJCfe29xx8j7v1pDL0QwbdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_E2u-m37KtXJoOySqa0', + ), + ( + 'https://www.youtube.com/s/player/8a8ac953/tv-player-es6.vflset/tv-player-es6.js', + '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA', + 'IAOAOq0QJ8wRAAgXmPlOPSBkkUs1bYFYlJCfe29xx8j7v1pDL0QwbdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_E2u-m37KtXJoOySqa0', + ), ] _NSIG_TESTS = [ @@ -272,7 +312,7 @@ _NSIG_TESTS = [ ), ( 'https://www.youtube.com/s/player/643afba4/player_ias.vflset/en_US/base.js', - 'W9HJZKktxuYoDTqW', 'larxUlagTRAcSw', + 'ir9-V6cdbCiyKxhr', '2PL7ZDYAALMfmA', ), ( 'https://www.youtube.com/s/player/363db69b/player_ias.vflset/en_US/base.js', @@ -286,6 +326,26 @@ _NSIG_TESTS = [ 'https://www.youtube.com/s/player/4fcd6e4a/tv-player-ias.vflset/tv-player-ias.js', 'o_L251jm8yhZkWtBW', 'lXoxI3XvToqn6A', ), + ( + 'https://www.youtube.com/s/player/20830619/tv-player-ias.vflset/tv-player-ias.js', + 'ir9-V6cdbCiyKxhr', '9YE85kNjZiS4', + ), + ( + 'https://www.youtube.com/s/player/20830619/player-plasma-ias-phone-en_US.vflset/base.js', + 'ir9-V6cdbCiyKxhr', '9YE85kNjZiS4', + ), + ( + 'https://www.youtube.com/s/player/20830619/player-plasma-ias-tablet-en_US.vflset/base.js', + 'ir9-V6cdbCiyKxhr', '9YE85kNjZiS4', + ), + ( + 'https://www.youtube.com/s/player/8a8ac953/player_ias_tce.vflset/en_US/base.js', + 'MiBYeXx_vRREbiCCmh', 'RtZYMVvmkE0JE', + ), + ( + 'https://www.youtube.com/s/player/8a8ac953/tv-player-es6.vflset/tv-player-es6.js', + 'MiBYeXx_vRREbiCCmh', 'RtZYMVvmkE0JE', + ), ] @@ -335,7 +395,7 @@ def t_factory(name, sig_func, url_pattern): test_id = re.sub(r'[/.-]', '_', m.group('id') or m.group('compat_id')) def test_func(self): - basename = 'player-{0}-{1}.js'.format(name, test_id) + basename = 'player-{0}.js'.format(test_id) fn = os.path.join(self.TESTDATA_DIR, basename) if not os.path.exists(fn): diff --git a/youtube_dl/YoutubeDL.py b/youtube_dl/YoutubeDL.py index 9e5620eef..8367b6e53 100755 --- a/youtube_dl/YoutubeDL.py +++ b/youtube_dl/YoutubeDL.py @@ -540,10 +540,14 @@ class YoutubeDL(object): """Print message to stdout if not in quiet mode.""" return self.to_stdout(message, skip_eol, check_quiet=True) - def _write_string(self, s, out=None): + def _write_string(self, s, out=None, only_once=False, _cache=set()): + if only_once and s in _cache: + return write_string(s, out=out, encoding=self.params.get('encoding')) + if only_once: + _cache.add(s) - def to_stdout(self, message, skip_eol=False, check_quiet=False): + def to_stdout(self, message, skip_eol=False, check_quiet=False, only_once=False): """Print message to stdout if not in quiet mode.""" if self.params.get('logger'): self.params['logger'].debug(message) @@ -552,9 +556,9 @@ class YoutubeDL(object): terminator = ['\n', ''][skip_eol] output = message + terminator - self._write_string(output, self._screen_file) + self._write_string(output, self._screen_file, only_once=only_once) - def to_stderr(self, message): + def to_stderr(self, message, only_once=False): """Print message to stderr.""" assert isinstance(message, compat_str) if self.params.get('logger'): @@ -562,7 +566,7 @@ class YoutubeDL(object): else: message = self._bidi_workaround(message) output = message + '\n' - self._write_string(output, self._err_file) + self._write_string(output, self._err_file, only_once=only_once) def to_console_title(self, message): if not self.params.get('consoletitle', False): @@ -641,18 +645,11 @@ class YoutubeDL(object): raise DownloadError(message, exc_info) self._download_retcode = 1 - def report_warning(self, message, only_once=False, _cache={}): + def report_warning(self, message, only_once=False): ''' Print the message to stderr, it will be prefixed with 'WARNING:' If stderr is a tty file the 'WARNING:' will be colored ''' - if only_once: - m_hash = hash((self, message)) - m_cnt = _cache.setdefault(m_hash, 0) - _cache[m_hash] = m_cnt + 1 - if m_cnt > 0: - return - if self.params.get('logger') is not None: self.params['logger'].warning(message) else: @@ -663,7 +660,7 @@ class YoutubeDL(object): else: _msg_header = 'WARNING:' warning_message = '%s %s' % (_msg_header, message) - self.to_stderr(warning_message) + self.to_stderr(warning_message, only_once=only_once) def report_error(self, message, *args, **kwargs): ''' @@ -677,6 +674,16 @@ class YoutubeDL(object): kwargs['message'] = '%s %s' % (_msg_header, message) self.trouble(*args, **kwargs) + def write_debug(self, message, only_once=False): + '''Log debug message or Print message to stderr''' + if not self.params.get('verbose', False): + return + message = '[debug] {0}'.format(message) + if self.params.get('logger'): + self.params['logger'].debug(message) + else: + self.to_stderr(message, only_once) + def report_unscoped_cookies(self, *args, **kwargs): # message=None, tb=False, is_error=False if len(args) <= 2: @@ -2514,7 +2521,7 @@ class YoutubeDL(object): self.get_encoding())) write_string(encoding_str, encoding=None) - writeln_debug = lambda *s: self._write_string('[debug] %s\n' % (''.join(s), )) + writeln_debug = lambda *s: self.write_debug(''.join(s)) writeln_debug('youtube-dl version ', __version__) if _LAZY_LOADER: writeln_debug('Lazy loading extractors enabled') diff --git a/youtube_dl/cache.py b/youtube_dl/cache.py index 54123da0e..ba12d0373 100644 --- a/youtube_dl/cache.py +++ b/youtube_dl/cache.py @@ -1,6 +1,6 @@ +# coding: utf-8 from __future__ import unicode_literals -import errno import json import os import re @@ -8,14 +8,17 @@ import shutil import traceback from .compat import ( + compat_contextlib_suppress, compat_getenv, compat_open as open, + compat_os_makedirs, ) from .utils import ( error_to_compat_str, + escape_rfc3986, expand_path, is_outdated_version, - try_get, + traverse_obj, write_json_file, ) from .version import __version__ @@ -30,23 +33,35 @@ class Cache(object): def __init__(self, ydl): self._ydl = ydl + def _write_debug(self, *args, **kwargs): + self._ydl.write_debug(*args, **kwargs) + + def _report_warning(self, *args, **kwargs): + self._ydl.report_warning(*args, **kwargs) + + def _to_screen(self, *args, **kwargs): + self._ydl.to_screen(*args, **kwargs) + + def _get_param(self, k, default=None): + return self._ydl.params.get(k, default) + def _get_root_dir(self): - res = self._ydl.params.get('cachedir') + res = self._get_param('cachedir') if res is None: cache_root = compat_getenv('XDG_CACHE_HOME', '~/.cache') res = os.path.join(cache_root, self._YTDL_DIR) return expand_path(res) def _get_cache_fn(self, section, key, dtype): - assert re.match(r'^[a-zA-Z0-9_.-]+$', section), \ + assert re.match(r'^[\w.-]+$', section), \ 'invalid section %r' % section - assert re.match(r'^[a-zA-Z0-9_.-]+$', key), 'invalid key %r' % key + key = escape_rfc3986(key, safe='').replace('%', ',') # encode non-ascii characters return os.path.join( self._get_root_dir(), section, '%s.%s' % (key, dtype)) @property def enabled(self): - return self._ydl.params.get('cachedir') is not False + return self._get_param('cachedir') is not False def store(self, section, key, data, dtype='json'): assert dtype in ('json',) @@ -56,61 +71,55 @@ class Cache(object): fn = self._get_cache_fn(section, key, dtype) try: - try: - os.makedirs(os.path.dirname(fn)) - except OSError as ose: - if ose.errno != errno.EEXIST: - raise + compat_os_makedirs(os.path.dirname(fn), exist_ok=True) + self._write_debug('Saving {section}.{key} to cache'.format(section=section, key=key)) write_json_file({self._VERSION_KEY: __version__, 'data': data}, fn) except Exception: tb = traceback.format_exc() - self._ydl.report_warning( - 'Writing cache to %r failed: %s' % (fn, tb)) + self._report_warning('Writing cache to {fn!r} failed: {tb}'.format(fn=fn, tb=tb)) def _validate(self, data, min_ver): - version = try_get(data, lambda x: x[self._VERSION_KEY]) + version = traverse_obj(data, self._VERSION_KEY) if not version: # Backward compatibility data, version = {'data': data}, self._DEFAULT_VERSION if not is_outdated_version(version, min_ver or '0', assume_new=False): return data['data'] - self._ydl.to_screen( - 'Discarding old cache from version {version} (needs {min_ver})'.format(**locals())) + self._write_debug('Discarding old cache from version {version} (needs {min_ver})'.format(version=version, min_ver=min_ver)) - def load(self, section, key, dtype='json', default=None, min_ver=None): + def load(self, section, key, dtype='json', default=None, **kw_min_ver): assert dtype in ('json',) + min_ver = kw_min_ver.get('min_ver') if not self.enabled: return default cache_fn = self._get_cache_fn(section, key, dtype) - try: + with compat_contextlib_suppress(IOError): # If no cache available try: - with open(cache_fn, 'r', encoding='utf-8') as cachef: + with open(cache_fn, encoding='utf-8') as cachef: + self._write_debug('Loading {section}.{key} from cache'.format(section=section, key=key), only_once=True) return self._validate(json.load(cachef), min_ver) - except ValueError: + except (ValueError, KeyError): try: file_size = os.path.getsize(cache_fn) except (OSError, IOError) as oe: file_size = error_to_compat_str(oe) - self._ydl.report_warning( - 'Cache retrieval from %s failed (%s)' % (cache_fn, file_size)) - except IOError: - pass # No cache available + self._report_warning('Cache retrieval from %s failed (%s)' % (cache_fn, file_size)) return default def remove(self): if not self.enabled: - self._ydl.to_screen('Cache is disabled (Did you combine --no-cache-dir and --rm-cache-dir?)') + self._to_screen('Cache is disabled (Did you combine --no-cache-dir and --rm-cache-dir?)') return cachedir = self._get_root_dir() if not any((term in cachedir) for term in ('cache', 'tmp')): - raise Exception('Not removing directory %s - this does not look like a cache dir' % cachedir) + raise Exception('Not removing directory %s - this does not look like a cache dir' % (cachedir,)) - self._ydl.to_screen( - 'Removing cache dir %s .' % cachedir, skip_eol=True) + self._to_screen( + 'Removing cache dir %s .' % (cachedir,), skip_eol=True, ), if os.path.exists(cachedir): - self._ydl.to_screen('.', skip_eol=True) + self._to_screen('.', skip_eol=True) shutil.rmtree(cachedir) - self._ydl.to_screen('.') + self._to_screen('.') diff --git a/youtube_dl/compat.py b/youtube_dl/compat.py index 8910a4dac..ebe22bdf9 100644 --- a/youtube_dl/compat.py +++ b/youtube_dl/compat.py @@ -2498,8 +2498,7 @@ try: from urllib.parse import urlencode as compat_urllib_parse_urlencode from urllib.parse import parse_qs as compat_parse_qs except ImportError: # Python 2 - _asciire = (compat_urllib_parse._asciire if hasattr(compat_urllib_parse, '_asciire') - else re.compile(r'([\x00-\x7f]+)')) + _asciire = getattr(compat_urllib_parse, '_asciire', None) or re.compile(r'([\x00-\x7f]+)') # HACK: The following are the correct unquote_to_bytes, unquote and unquote_plus # implementations from cpython 3.4.3's stdlib. Python 2's version @@ -2567,24 +2566,21 @@ except ImportError: # Python 2 # Possible solutions are to either port it from python 3 with all # the friends or manually ensure input query contains only byte strings. # We will stick with latter thus recursively encoding the whole query. - def compat_urllib_parse_urlencode(query, doseq=0, encoding='utf-8'): + def compat_urllib_parse_urlencode(query, doseq=0, safe='', encoding='utf-8', errors='strict'): + def encode_elem(e): if isinstance(e, dict): e = encode_dict(e) elif isinstance(e, (list, tuple,)): - list_e = encode_list(e) - e = tuple(list_e) if isinstance(e, tuple) else list_e + e = type(e)(encode_elem(el) for el in e) elif isinstance(e, compat_str): - e = e.encode(encoding) + e = e.encode(encoding, errors) return e def encode_dict(d): - return dict((encode_elem(k), encode_elem(v)) for k, v in d.items()) - - def encode_list(l): - return [encode_elem(e) for e in l] + return tuple((encode_elem(k), encode_elem(v)) for k, v in d.items()) - return compat_urllib_parse._urlencode(encode_elem(query), doseq=doseq) + return compat_urllib_parse._urlencode(encode_elem(query), doseq=doseq).decode('ascii') # HACK: The following is the correct parse_qs implementation from cpython 3's stdlib. # Python 2's version is apparently totally broken @@ -2639,6 +2635,57 @@ except ImportError: # Python 2 ('parse_qs', compat_parse_qs)): setattr(compat_urllib_parse, name, fix) + try: + all(chr(i) in b'' for i in range(256)) + except TypeError: + # not all chr(i) are str: patch Python2 quote + + _safemaps = getattr(compat_urllib_parse, '_safemaps', {}) + _always_safe = frozenset(compat_urllib_parse.always_safe) + + def _quote(s, safe='/'): + """quote('abc def') -> 'abc%20def'""" + + if not s and s is not None: # fast path + return s + safe = frozenset(safe) + cachekey = (safe, _always_safe) + try: + safe_map = _safemaps[cachekey] + except KeyError: + safe = _always_safe | safe + safe_map = {} + for i in range(256): + c = chr(i) + safe_map[c] = ( + c if (i < 128 and c in safe) + else b'%{0:02X}'.format(i)) + _safemaps[cachekey] = safe_map + + if safe.issuperset(s): + return s + return ''.join(safe_map[c] for c in s) + + # linked code + def _quote_plus(s, safe=''): + return ( + _quote(s, safe + b' ').replace(b' ', b'+') if b' ' in s + else _quote(s, safe)) + + # linked code + def _urlcleanup(): + if compat_urllib_parse._urlopener: + compat_urllib_parse._urlopener.cleanup() + _safemaps.clear() + compat_urllib_parse.ftpcache.clear() + + for name, fix in ( + ('quote', _quote), + ('quote_plus', _quote_plus), + ('urlcleanup', _urlcleanup)): + setattr(compat_urllib_parse, '_' + name, getattr(compat_urllib_parse, name)) + setattr(compat_urllib_parse, name, fix) + compat_urllib_parse_parse_qs = compat_parse_qs @@ -3120,6 +3167,21 @@ else: compat_os_path_expanduser = compat_expanduser +# compat_os_makedirs +try: + os.makedirs('.', exist_ok=True) + compat_os_makedirs = os.makedirs +except TypeError: # < Py3.2 + from errno import EEXIST as _errno_EEXIST + + def compat_os_makedirs(name, mode=0o777, exist_ok=False): + try: + return os.makedirs(name, mode=mode) + except OSError as ose: + if not (exist_ok and ose.errno == _errno_EEXIST): + raise + + # compat_os_path_realpath if compat_os_name == 'nt' and sys.version_info < (3, 8): # os.path.realpath on Windows does not follow symbolic links @@ -3637,6 +3699,7 @@ __all__ = [ 'compat_numeric_types', 'compat_open', 'compat_ord', + 'compat_os_makedirs', 'compat_os_name', 'compat_os_path_expanduser', 'compat_os_path_realpath', diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py index cb67b976d..a64fcfccc 100644 --- a/youtube_dl/extractor/common.py +++ b/youtube_dl/extractor/common.py @@ -505,7 +505,7 @@ class InfoExtractor(object): if not self._x_forwarded_for_ip: # Geo bypass mechanism is explicitly disabled by user - if not self._downloader.params.get('geo_bypass', True): + if not self.get_param('geo_bypass', True): return if not geo_bypass_context: @@ -527,7 +527,7 @@ class InfoExtractor(object): # Explicit IP block specified by user, use it right away # regardless of whether extractor is geo bypassable or not - ip_block = self._downloader.params.get('geo_bypass_ip_block', None) + ip_block = self.get_param('geo_bypass_ip_block', None) # Otherwise use random IP block from geo bypass context but only # if extractor is known as geo bypassable @@ -538,8 +538,8 @@ class InfoExtractor(object): if ip_block: self._x_forwarded_for_ip = GeoUtils.random_ipv4(ip_block) - if self._downloader.params.get('verbose', False): - self._downloader.to_screen( + if self.get_param('verbose', False): + self.to_screen( '[debug] Using fake IP %s as X-Forwarded-For.' % self._x_forwarded_for_ip) return @@ -548,7 +548,7 @@ class InfoExtractor(object): # Explicit country code specified by user, use it right away # regardless of whether extractor is geo bypassable or not - country = self._downloader.params.get('geo_bypass_country', None) + country = self.get_param('geo_bypass_country', None) # Otherwise use random country code from geo bypass context but # only if extractor is known as geo bypassable @@ -559,8 +559,8 @@ class InfoExtractor(object): if country: self._x_forwarded_for_ip = GeoUtils.random_ipv4(country) - if self._downloader.params.get('verbose', False): - self._downloader.to_screen( + if self.get_param('verbose', False): + self.to_screen( '[debug] Using fake IP %s (%s) as X-Forwarded-For.' % (self._x_forwarded_for_ip, country.upper())) @@ -586,9 +586,9 @@ class InfoExtractor(object): raise ExtractorError('An extractor error has occurred.', cause=e) def __maybe_fake_ip_and_retry(self, countries): - if (not self._downloader.params.get('geo_bypass_country', None) + if (not self.get_param('geo_bypass_country', None) and self._GEO_BYPASS - and self._downloader.params.get('geo_bypass', True) + and self.get_param('geo_bypass', True) and not self._x_forwarded_for_ip and countries): country_code = random.choice(countries) @@ -698,7 +698,7 @@ class InfoExtractor(object): if fatal: raise ExtractorError(errmsg, sys.exc_info()[2], cause=err) else: - self._downloader.report_warning(errmsg) + self.report_warning(errmsg) return False def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None): @@ -770,11 +770,11 @@ class InfoExtractor(object): webpage_bytes = prefix + webpage_bytes if not encoding: encoding = self._guess_encoding_from_content(content_type, webpage_bytes) - if self._downloader.params.get('dump_intermediate_pages', False): + if self.get_param('dump_intermediate_pages', False): self.to_screen('Dumping request to ' + urlh.geturl()) dump = base64.b64encode(webpage_bytes).decode('ascii') - self._downloader.to_screen(dump) - if self._downloader.params.get('write_pages', False): + self.to_screen(dump) + if self.get_param('write_pages', False): basen = '%s_%s' % (video_id, urlh.geturl()) if len(basen) > 240: h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest() @@ -976,19 +976,9 @@ class InfoExtractor(object): """Print msg to screen, prefixing it with '[ie_name]'""" self._downloader.to_screen(self.__ie_msg(msg)) - def write_debug(self, msg, only_once=False, _cache=[]): + def write_debug(self, msg, only_once=False): '''Log debug message or Print message to stderr''' - if not self.get_param('verbose', False): - return - message = '[debug] ' + self.__ie_msg(msg) - logger = self.get_param('logger') - if logger: - logger.debug(message) - else: - if only_once and hash(message) in _cache: - return - self._downloader.to_stderr(message) - _cache.append(hash(message)) + self._downloader.write_debug(self.__ie_msg(msg), only_once=only_once) # name, default=None, *args, **kwargs def get_param(self, name, *args, **kwargs): @@ -1084,7 +1074,7 @@ class InfoExtractor(object): if mobj: break - if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty(): + if not self.get_param('no_color') and compat_os_name != 'nt' and sys.stderr.isatty(): _name = '\033[0;34m%s\033[0m' % name else: _name = name @@ -1102,7 +1092,7 @@ class InfoExtractor(object): elif fatal: raise RegexNotFoundError('Unable to extract %s' % _name) else: - self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message()) + self.report_warning('unable to extract %s' % _name + bug_reports_message()) return None def _search_json(self, start_pattern, string, name, video_id, **kwargs): @@ -1172,7 +1162,7 @@ class InfoExtractor(object): username = None password = None - if self._downloader.params.get('usenetrc', False): + if self.get_param('usenetrc', False): try: netrc_machine = netrc_machine or self._NETRC_MACHINE info = netrc.netrc().authenticators(netrc_machine) @@ -1183,7 +1173,7 @@ class InfoExtractor(object): raise netrc.NetrcParseError( 'No authenticators for %s' % netrc_machine) except (AttributeError, IOError, netrc.NetrcParseError) as err: - self._downloader.report_warning( + self.report_warning( 'parsing .netrc: %s' % error_to_compat_str(err)) return username, password @@ -1220,10 +1210,10 @@ class InfoExtractor(object): """ if self._downloader is None: return None - downloader_params = self._downloader.params - if downloader_params.get('twofactor') is not None: - return downloader_params['twofactor'] + twofactor = self.get_param('twofactor') + if twofactor is not None: + return twofactor return compat_getpass('Type %s and press [Return]: ' % note) @@ -1358,7 +1348,7 @@ class InfoExtractor(object): elif fatal: raise RegexNotFoundError('Unable to extract JSON-LD') else: - self._downloader.report_warning('unable to extract JSON-LD %s' % bug_reports_message()) + self.report_warning('unable to extract JSON-LD %s' % bug_reports_message()) return {} def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None): @@ -1589,7 +1579,7 @@ class InfoExtractor(object): if f.get('vcodec') == 'none': # audio only preference -= 50 - if self._downloader.params.get('prefer_free_formats'): + if self.get_param('prefer_free_formats'): ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus'] else: ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a'] @@ -1601,7 +1591,7 @@ class InfoExtractor(object): else: if f.get('acodec') == 'none': # video only preference -= 40 - if self._downloader.params.get('prefer_free_formats'): + if self.get_param('prefer_free_formats'): ORDER = ['flv', 'mp4', 'webm'] else: ORDER = ['webm', 'flv', 'mp4'] @@ -1667,7 +1657,7 @@ class InfoExtractor(object): """ Either "http:" or "https:", depending on the user's preferences """ return ( 'http:' - if self._downloader.params.get('prefer_insecure', False) + if self.get_param('prefer_insecure', False) else 'https:') def _proto_relative_url(self, url, scheme=None): @@ -3199,7 +3189,7 @@ class InfoExtractor(object): if fatal: raise ExtractorError(msg) else: - self._downloader.report_warning(msg) + self.report_warning(msg) return res def _float(self, v, name, fatal=False, **kwargs): @@ -3209,7 +3199,7 @@ class InfoExtractor(object): if fatal: raise ExtractorError(msg) else: - self._downloader.report_warning(msg) + self.report_warning(msg) return res def _set_cookie(self, domain, name, value, expire_time=None, port=None, @@ -3218,12 +3208,12 @@ class InfoExtractor(object): 0, name, value, port, port is not None, domain, True, domain.startswith('.'), path, True, secure, expire_time, discard, None, None, rest) - self._downloader.cookiejar.set_cookie(cookie) + self.cookiejar.set_cookie(cookie) def _get_cookies(self, url): """ Return a compat_cookies_SimpleCookie with the cookies for the url """ req = sanitized_Request(url) - self._downloader.cookiejar.add_cookie_header(req) + self.cookiejar.add_cookie_header(req) return compat_cookies_SimpleCookie(req.get_header('Cookie')) def _apply_first_set_cookie_header(self, url_handle, cookie): @@ -3283,8 +3273,8 @@ class InfoExtractor(object): return not any_restricted def extract_subtitles(self, *args, **kwargs): - if (self._downloader.params.get('writesubtitles', False) - or self._downloader.params.get('listsubtitles')): + if (self.get_param('writesubtitles', False) + or self.get_param('listsubtitles')): return self._get_subtitles(*args, **kwargs) return {} @@ -3305,7 +3295,11 @@ class InfoExtractor(object): """ Merge subtitle dictionaries, language by language. """ # ..., * , target=None - target = kwargs.get('target') or dict(subtitle_dict1) + target = kwargs.get('target') + if target is None: + target = dict(subtitle_dict1) + else: + subtitle_dicts = (subtitle_dict1,) + subtitle_dicts for subtitle_dict in subtitle_dicts: for lang in subtitle_dict: @@ -3313,8 +3307,8 @@ class InfoExtractor(object): return target def extract_automatic_captions(self, *args, **kwargs): - if (self._downloader.params.get('writeautomaticsub', False) - or self._downloader.params.get('listsubtitles')): + if (self.get_param('writeautomaticsub', False) + or self.get_param('listsubtitles')): return self._get_automatic_captions(*args, **kwargs) return {} @@ -3322,9 +3316,9 @@ class InfoExtractor(object): raise NotImplementedError('This method must be implemented by subclasses') def mark_watched(self, *args, **kwargs): - if (self._downloader.params.get('mark_watched', False) + if (self.get_param('mark_watched', False) and (self._get_login_info()[0] is not None - or self._downloader.params.get('cookiefile') is not None)): + or self.get_param('cookiefile') is not None)): self._mark_watched(*args, **kwargs) def _mark_watched(self, *args, **kwargs): @@ -3332,7 +3326,7 @@ class InfoExtractor(object): def geo_verification_headers(self): headers = {} - geo_verification_proxy = self._downloader.params.get('geo_verification_proxy') + geo_verification_proxy = self.get_param('geo_verification_proxy') if geo_verification_proxy: headers['Ytdl-request-proxy'] = geo_verification_proxy return headers diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index ce97fd75b..54073ef86 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -342,14 +342,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor): if not self._login(): return - _DEFAULT_API_DATA = { - 'context': { - 'client': { - 'clientName': 'WEB', - 'clientVersion': '2.20201021.03.00', - }, - }, - } + _DEFAULT_API_DATA = {'context': _INNERTUBE_CLIENTS['web']['INNERTUBE_CONTEXT']} _YT_INITIAL_DATA_RE = r'(?:window\s*\[\s*["\']ytInitialData["\']\s*\]|ytInitialData)\s*=\s*({.+?})\s*;' _YT_INITIAL_PLAYER_RESPONSE_RE = r'ytInitialPlayerResponse\s*=\s*({.+?})\s*;' @@ -497,11 +490,15 @@ class YoutubeBaseInfoExtractor(InfoExtractor): data['params'] = params for page_num in itertools.count(1): search = self._download_json( - 'https://www.youtube.com/youtubei/v1/search?key=AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8', + 'https://www.youtube.com/youtubei/v1/search', video_id='query "%s"' % query, note='Downloading page %s' % page_num, errnote='Unable to download API page', fatal=False, data=json.dumps(data).encode('utf8'), + query={ + # 'key': 'AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8', + 'prettyPrint': 'false', + }, headers={'content-type': 'application/json'}) if not search: break @@ -1655,7 +1652,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): assert os.path.basename(func_id) == func_id self.write_debug('Extracting signature function {0}'.format(func_id)) - cache_spec, code = self.cache.load('youtube-sigfuncs', func_id), None + cache_spec, code = self.cache.load('youtube-sigfuncs', func_id, min_ver='2025.04.07'), None if not cache_spec: code = self._load_player(video_id, player_url, player_id) @@ -1816,6 +1813,10 @@ class YoutubeIE(YoutubeBaseInfoExtractor): return ret def _extract_n_function_name(self, jscode): + func_name, idx = None, None + # these special cases are redundant and probably obsolete (2025-04): + # they make the tests run ~10% faster without fallback warnings + r""" func_name, idx = self._search_regex( # (y=NuD(),Mw(k),q=k.Z[y]||null)&&(q=narray[idx](q),k.set(y,q),k.V||NuD(''))}}; # (R="nn"[+J.Z],mW(J),N=J.K[R]||null)&&(N=narray[idx](N),J.set(R,N))}}; @@ -1842,9 +1843,28 @@ class YoutubeIE(YoutubeBaseInfoExtractor): \(\s*[\w$]+\s*\) ''', jscode, 'Initial JS player n function name', group=('nfunc', 'idx'), default=(None, None)) + """ + + if not func_name: + # nfunc=function(x){...}|function nfunc(x); ... + # ... var y=[nfunc]|y[idx]=nfunc); + # obvious REs hang, so use a two-stage tactic + for m in re.finditer(r'''(?x) + [\n;]var\s(?:(?:(?!,).)+,|\s)*?(?!\d)[\w$]+(?:\[(?P<idx>\d+)\])?\s*=\s* + (?(idx)|\[\s*)(?P<nfunc>(?!\d)[\w$]+)(?(idx)|\s*\]) + \s*?[;\n] + ''', jscode): + func_name = self._search_regex( + r'[;,]\s*(function\s+)?({0})(?(1)|\s*=\s*function)\s*\((?!\d)[\w$]+\)\s*\{1}(?!\s*return\s)'.format( + re.escape(m.group('nfunc')), '{'), + jscode, 'Initial JS player n function name (2)', group=2, default=None) + if func_name: + idx = m.group('idx') + break + # thx bashonly: yt-dlp/yt-dlp/pull/10611 if not func_name: - self.report_warning('Falling back to generic n function search') + self.report_warning('Falling back to generic n function search', only_once=True) return self._search_regex( r'''(?xs) (?:(?<=[^\w$])|^) # instead of \b, which ignores $ @@ -1858,14 +1878,14 @@ class YoutubeIE(YoutubeBaseInfoExtractor): return func_name return self._search_json( - r'var\s+{0}\s*='.format(re.escape(func_name)), jscode, + r'(?<![\w-])var\s(?:(?:(?!,).)+,|\s)*?{0}\s*='.format(re.escape(func_name)), jscode, 'Initial JS player n function list ({0}.{1})'.format(func_name, idx), - func_name, contains_pattern=r'\[[\s\S]+\]', end_pattern='[,;]', + func_name, contains_pattern=r'\[.+\]', end_pattern='[,;]', transform_source=js_to_json)[int(idx)] def _extract_n_function_code(self, video_id, player_url): player_id = self._extract_player_info(player_url) - func_code = self.cache.load('youtube-nsig', player_id) + func_code = self.cache.load('youtube-nsig', player_id, min_ver='2025.04.07') jscode = func_code or self._load_player(video_id, player_url) jsi = JSInterpreter(jscode) @@ -3339,6 +3359,20 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): 'thumbnailViewModel', 'image'), final_key='sources'), }) + def _extract_shorts_lockup_view_model(self, view_model): + content_id = traverse_obj(view_model, ( + 'onTap', 'innertubeCommand', 'reelWatchEndpoint', 'videoId', + T(lambda v: v if YoutubeIE.suitable(v) else None))) + if not content_id: + return + return merge_dicts(self.url_result( + content_id, ie=YoutubeIE.ie_key(), video_id=content_id), { + 'title': traverse_obj(view_model, ( + 'overlayMetadata', 'primaryText', 'content', T(compat_str))), + 'thumbnails': self._extract_thumbnails( + view_model, 'thumbnail', final_key='sources'), + }) + def _video_entry(self, video_renderer): video_id = video_renderer.get('videoId') if video_id: @@ -3385,10 +3419,9 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): yield entry def _rich_grid_entries(self, contents): - for content in contents: - content = traverse_obj( - content, ('richItemRenderer', 'content'), - expected_type=dict) or {} + for content in traverse_obj( + contents, (Ellipsis, 'richItemRenderer', 'content'), + expected_type=dict): video_renderer = traverse_obj( content, 'videoRenderer', 'reelItemRenderer', expected_type=dict) @@ -3396,6 +3429,12 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): entry = self._video_entry(video_renderer) if entry: yield entry + # shorts item + shorts_lockup_view_model = content.get('shortsLockupViewModel') + if shorts_lockup_view_model: + entry = self._extract_shorts_lockup_view_model(shorts_lockup_view_model) + if entry: + yield entry # playlist renderer = traverse_obj( content, 'playlistRenderer', expected_type=dict) or {} @@ -3434,23 +3473,15 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): next_continuation = cls._extract_next_continuation_data(renderer) if next_continuation: return next_continuation - contents = [] - for key in ('contents', 'items'): - contents.extend(try_get(renderer, lambda x: x[key], list) or []) - for content in contents: - if not isinstance(content, dict): - continue - continuation_ep = try_get( - content, lambda x: x['continuationItemRenderer']['continuationEndpoint'], - dict) - if not continuation_ep: - continue - continuation = try_get( - continuation_ep, lambda x: x['continuationCommand']['token'], compat_str) + for command in traverse_obj(renderer, ( + ('contents', 'items', 'rows'), Ellipsis, 'continuationItemRenderer', + ('continuationEndpoint', ('button', 'buttonRenderer', 'command')), + (('commandExecutorCommand', 'commands', Ellipsis), None), T(dict))): + continuation = traverse_obj(command, ('continuationCommand', 'token', T(compat_str))) if not continuation: continue - ctp = continuation_ep.get('clickTrackingParams') - return YoutubeTabIE._build_continuation_query(continuation, ctp) + ctp = command.get('clickTrackingParams') + return cls._build_continuation_query(continuation, ctp) def _entries(self, tab, item_id, webpage): tab_content = try_get(tab, lambda x: x['content'], dict) @@ -3499,6 +3530,13 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): entry = self._video_entry(renderer) if entry: yield entry + renderer = isr_content.get('richGridRenderer') + if renderer: + for from_ in self._rich_grid_entries( + traverse_obj(renderer, ('contents', Ellipsis, T(dict)))): + yield from_ + continuation = self._extract_continuation(renderer) + continue if not continuation: continuation = self._extract_continuation(is_renderer) @@ -3508,8 +3546,9 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): rich_grid_renderer = tab_content.get('richGridRenderer') if not rich_grid_renderer: return - for entry in self._rich_grid_entries(rich_grid_renderer.get('contents') or []): - yield entry + for from_ in self._rich_grid_entries( + traverse_obj(rich_grid_renderer, ('contents', Ellipsis, T(dict)))): + yield from_ continuation = self._extract_continuation(rich_grid_renderer) @@ -3555,8 +3594,12 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): # Downloading page may result in intermittent 5xx HTTP error # that is usually worked around with a retry response = self._download_json( - 'https://www.youtube.com/youtubei/v1/browse?key=AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8', + 'https://www.youtube.com/youtubei/v1/browse', None, 'Downloading page %d%s' % (page_num, ' (retry #%d)' % count if count else ''), + query={ + # 'key': 'AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8', + 'prettyPrint': 'false', + }, headers=headers, data=json.dumps(data).encode('utf8')) break except ExtractorError as e: diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index f0c4fa928..7630e2099 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -240,7 +240,7 @@ def _js_ternary(cndn, if_true=True, if_false=False): def _js_unary_op(op): @wraps_op(op) - def wrapped(_, a): + def wrapped(a, _): return op(a) return wrapped @@ -283,17 +283,6 @@ _OPERATORS = ( ('**', _js_exp), ) -_COMP_OPERATORS = ( - ('===', _js_id_op(operator.is_)), - ('!==', _js_id_op(operator.is_not)), - ('==', _js_eq), - ('!=', _js_neq), - ('<=', _js_comp_op(operator.le)), - ('>=', _js_comp_op(operator.ge)), - ('<', _js_comp_op(operator.lt)), - ('>', _js_comp_op(operator.gt)), -) - _LOG_OPERATORS = ( ('|', _js_bit_op(operator.or_)), ('^', _js_bit_op(operator.xor)), @@ -310,13 +299,27 @@ _SC_OPERATORS = ( _UNARY_OPERATORS_X = ( ('void', _js_unary_op(lambda _: JS_Undefined)), ('typeof', _js_unary_op(_js_typeof)), + # avoid functools.partial here since Py2 update_wrapper(partial) -> no __module__ + ('!', _js_unary_op(lambda x: _js_ternary(x, if_true=False, if_false=True))), ) -_OPERATOR_RE = '|'.join(map(lambda x: re.escape(x[0]), _OPERATORS + _LOG_OPERATORS)) +_COMP_OPERATORS = ( + ('===', _js_id_op(operator.is_)), + ('!==', _js_id_op(operator.is_not)), + ('==', _js_eq), + ('!=', _js_neq), + ('<=', _js_comp_op(operator.le)), + ('>=', _js_comp_op(operator.ge)), + ('<', _js_comp_op(operator.lt)), + ('>', _js_comp_op(operator.gt)), +) + +_OPERATOR_RE = '|'.join(map(lambda x: re.escape(x[0]), _OPERATORS + _LOG_OPERATORS + _SC_OPERATORS)) _NAME_RE = r'[a-zA-Z_$][\w$]*' _MATCHING_PARENS = dict(zip(*zip('()', '{}', '[]'))) _QUOTES = '\'"/' +_NESTED_BRACKETS = r'[^[\]]+(?:\[[^[\]]+(?:\[[^\]]+\])?\])?' class JS_Break(ExtractorError): @@ -353,7 +356,7 @@ class LocalNameSpace(ChainMap): raise NotImplementedError('Deleting is not supported') def __repr__(self): - return 'LocalNameSpace%s' % (self.maps, ) + return 'LocalNameSpace({0!r})'.format(self.maps) class Debugger(object): @@ -374,6 +377,9 @@ class Debugger(object): @classmethod def wrap_interpreter(cls, f): + if not cls.ENABLED: + return f + @wraps(f) def interpret_statement(self, stmt, local_vars, allow_recursion, *args, **kwargs): if cls.ENABLED and stmt.strip(): @@ -414,7 +420,17 @@ class JSInterpreter(object): msg = '{0} in: {1!r:.100}'.format(msg.rstrip(), expr) super(JSInterpreter.Exception, self).__init__(msg, *args, **kwargs) - class JS_RegExp(object): + class JS_Object(object): + def __getitem__(self, key): + if hasattr(self, key): + return getattr(self, key) + raise KeyError(key) + + def dump(self): + """Serialise the instance""" + raise NotImplementedError + + class JS_RegExp(JS_Object): RE_FLAGS = { # special knowledge: Python's re flags are bitmask values, current max 128 # invent new bitmask values well above that for literal parsing @@ -435,16 +451,24 @@ class JSInterpreter(object): def __init__(self, pattern_txt, flags=0): if isinstance(flags, compat_str): flags, _ = self.regex_flags(flags) - # First, avoid https://github.com/python/cpython/issues/74534 self.__self = None pattern_txt = str_or_none(pattern_txt) or '(?:)' - self.__pattern_txt = pattern_txt.replace('[[', r'[\[') + # escape unintended embedded flags + pattern_txt = re.sub( + r'(\(\?)([aiLmsux]*)(-[imsx]+:|(?<!\?)\))', + lambda m: ''.join( + (re.escape(m.group(1)), m.group(2), re.escape(m.group(3))) + if m.group(3) == ')' + else ('(?:', m.group(2), m.group(3))), + pattern_txt) + # Avoid https://github.com/python/cpython/issues/74534 + self.source = pattern_txt.replace('[[', r'[\[') self.__flags = flags def __instantiate(self): if self.__self: return - self.__self = re.compile(self.__pattern_txt, self.__flags) + self.__self = re.compile(self.source, self.__flags) # Thx: https://stackoverflow.com/questions/44773522/setattr-on-python2-sre-sre-pattern for name in dir(self.__self): # Only these? Obviously __class__, __init__. @@ -452,16 +476,15 @@ class JSInterpreter(object): # that can't be setattr'd but also can't need to be copied. if name in ('__class__', '__init__', '__weakref__'): continue - setattr(self, name, getattr(self.__self, name)) + if name == 'flags': + setattr(self, name, getattr(self.__self, name, self.__flags)) + else: + setattr(self, name, getattr(self.__self, name)) def __getattr__(self, name): self.__instantiate() - # make Py 2.6 conform to its lying documentation - if name == 'flags': - self.flags = self.__flags - return self.flags - elif name == 'pattern': - self.pattern = self.__pattern_txt + if name == 'pattern': + self.pattern = self.source return self.pattern elif hasattr(self.__self, name): v = getattr(self.__self, name) @@ -469,6 +492,26 @@ class JSInterpreter(object): return v elif name in ('groupindex', 'groups'): return 0 if name == 'groupindex' else {} + else: + flag_attrs = ( # order by 2nd elt + ('hasIndices', 'd'), + ('global', 'g'), + ('ignoreCase', 'i'), + ('multiline', 'm'), + ('dotAll', 's'), + ('unicode', 'u'), + ('unicodeSets', 'v'), + ('sticky', 'y'), + ) + for k, c in flag_attrs: + if name == k: + return bool(self.RE_FLAGS[c] & self.__flags) + else: + if name == 'flags': + return ''.join( + (c if self.RE_FLAGS[c] & self.__flags else '') + for _, c in flag_attrs) + raise AttributeError('{0} has no attribute named {1}'.format(self, name)) @classmethod @@ -482,7 +525,16 @@ class JSInterpreter(object): flags |= cls.RE_FLAGS[ch] return flags, expr[idx + 1:] - class JS_Date(object): + def dump(self): + return '(/{0}/{1})'.format( + re.sub(r'(?<!\\)/', r'\/', self.source), + self.flags) + + @staticmethod + def escape(string_): + return re.escape(string_) + + class JS_Date(JS_Object): _t = None @staticmethod @@ -549,6 +601,9 @@ class JSInterpreter(object): def valueOf(self): return _NaN if self._t is None else self._t + def dump(self): + return '(new Date({0}))'.format(self.toString()) + @classmethod def __op_chars(cls): op_chars = set(';,[') @@ -652,6 +707,68 @@ class JSInterpreter(object): _SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS, _UNARY_OPERATORS_X)) return _cached + def _separate_at_op(self, expr, max_split=None): + + for op, _ in self._all_operators(): + # hackety: </> have higher priority than <</>>, but don't confuse them + skip_delim = (op + op) if op in '<>*?' else None + if op == '?': + skip_delim = (skip_delim, '?.') + separated = list(self._separate(expr, op, skip_delims=skip_delim)) + if len(separated) < 2: + continue + + right_expr = separated.pop() + # handle operators that are both unary and binary, minimal BODMAS + if op in ('+', '-'): + # simplify/adjust consecutive instances of these operators + undone = 0 + separated = [s.strip() for s in separated] + while len(separated) > 1 and not separated[-1]: + undone += 1 + separated.pop() + if op == '-' and undone % 2 != 0: + right_expr = op + right_expr + elif op == '+': + while len(separated) > 1 and set(separated[-1]) <= self.OP_CHARS: + right_expr = separated.pop() + right_expr + if separated[-1][-1:] in self.OP_CHARS: + right_expr = separated.pop() + right_expr + # hanging op at end of left => unary + (strip) or - (push right) + separated.append(right_expr) + dm_ops = ('*', '%', '/', '**') + dm_chars = set(''.join(dm_ops)) + + def yield_terms(s): + skip = False + for i, term in enumerate(s[:-1]): + if skip: + skip = False + continue + if not (dm_chars & set(term)): + yield term + continue + for dm_op in dm_ops: + bodmas = list(self._separate(term, dm_op, skip_delims=skip_delim)) + if len(bodmas) > 1 and not bodmas[-1].strip(): + bodmas[-1] = (op if op == '-' else '') + s[i + 1] + yield dm_op.join(bodmas) + skip = True + break + else: + if term: + yield term + + if not skip and s[-1]: + yield s[-1] + + separated = list(yield_terms(separated)) + right_expr = separated.pop() if len(separated) > 1 else None + expr = op.join(separated) + if right_expr is None: + continue + return op, separated, right_expr + def _operator(self, op, left_val, right_expr, expr, local_vars, allow_recursion): if op in ('||', '&&'): if (op == '&&') ^ _js_ternary(left_val): @@ -662,7 +779,7 @@ class JSInterpreter(object): elif op == '?': right_expr = _js_ternary(left_val, *self._separate(right_expr, ':', 1)) - right_val = self.interpret_expression(right_expr, local_vars, allow_recursion) + right_val = self.interpret_expression(right_expr, local_vars, allow_recursion) if right_expr else left_val opfunc = op and next((v for k, v in self._all_operators() if k == op), None) if not opfunc: return right_val @@ -686,6 +803,8 @@ class JSInterpreter(object): raise self.Exception('Cannot get index {idx!r:.100}'.format(**locals()), expr=repr(obj), cause=e) def _dump(self, obj, namespace): + if obj is JS_Undefined: + return 'undefined' try: return json.dumps(obj) except TypeError: @@ -705,51 +824,9 @@ class JSInterpreter(object): _FINALLY_RE = re.compile(r'finally\s*\{') _SWITCH_RE = re.compile(r'switch\s*\(') - def handle_operators(self, expr, local_vars, allow_recursion): - - for op, _ in self._all_operators(): - # hackety: </> have higher priority than <</>>, but don't confuse them - skip_delim = (op + op) if op in '<>*?' else None - if op == '?': - skip_delim = (skip_delim, '?.') - separated = list(self._separate(expr, op, skip_delims=skip_delim)) - if len(separated) < 2: - continue - - right_expr = separated.pop() - # handle operators that are both unary and binary, minimal BODMAS - if op in ('+', '-'): - # simplify/adjust consecutive instances of these operators - undone = 0 - separated = [s.strip() for s in separated] - while len(separated) > 1 and not separated[-1]: - undone += 1 - separated.pop() - if op == '-' and undone % 2 != 0: - right_expr = op + right_expr - elif op == '+': - while len(separated) > 1 and set(separated[-1]) <= self.OP_CHARS: - right_expr = separated.pop() + right_expr - if separated[-1][-1:] in self.OP_CHARS: - right_expr = separated.pop() + right_expr - # hanging op at end of left => unary + (strip) or - (push right) - left_val = separated[-1] if separated else '' - for dm_op in ('*', '%', '/', '**'): - bodmas = tuple(self._separate(left_val, dm_op, skip_delims=skip_delim)) - if len(bodmas) > 1 and not bodmas[-1].strip(): - expr = op.join(separated) + op + right_expr - if len(separated) > 1: - separated.pop() - right_expr = op.join((left_val, right_expr)) - else: - separated = [op.join((left_val, right_expr))] - right_expr = None - break - if right_expr is None: - continue - - left_val = self.interpret_expression(op.join(separated), local_vars, allow_recursion) - return self._operator(op, left_val, right_expr, expr, local_vars, allow_recursion), True + def _eval_operator(self, op, left_expr, right_expr, expr, local_vars, allow_recursion): + left_val = self.interpret_expression(left_expr, local_vars, allow_recursion) + return self._operator(op, left_val, right_expr, expr, local_vars, allow_recursion) @Debugger.wrap_interpreter def interpret_statement(self, stmt, local_vars, allow_recursion=100): @@ -805,15 +882,19 @@ class JSInterpreter(object): else: raise self.Exception('Unsupported object {obj:.100}'.format(**locals()), expr=expr) + # apply unary operators (see new above) for op, _ in _UNARY_OPERATORS_X: if not expr.startswith(op): continue operand = expr[len(op):] - if not operand or operand[0] != ' ': + if not operand or (op.isalpha() and operand[0] != ' '): continue - op_result = self.handle_operators(expr, local_vars, allow_recursion) - if op_result: - return op_result[0], should_return + separated = self._separate_at_op(operand, max_split=1) + if separated: + next_op, separated, right_expr = separated + separated.append(right_expr) + operand = next_op.join(separated) + return self._eval_operator(op, operand, '', expr, local_vars, allow_recursion), should_return if expr.startswith('{'): inner, outer = self._separate_at_paren(expr) @@ -1008,15 +1089,18 @@ class JSInterpreter(object): m = re.match(r'''(?x) (?P<assign> - (?P<out>{_NAME_RE})(?:\[(?P<out_idx>(?:.+?\]\s*\[)*.+?)\])?\s* + (?P<out>{_NAME_RE})(?P<out_idx>(?:\[{_NESTED_BRACKETS}\])+)?\s* (?P<op>{_OPERATOR_RE})? =(?!=)(?P<expr>.*)$ )|(?P<return> (?!if|return|true|false|null|undefined|NaN|Infinity)(?P<name>{_NAME_RE})$ - )|(?P<indexing> - (?P<in>{_NAME_RE})\[(?P<in_idx>(?:.+?\]\s*\[)*.+?)\]$ )|(?P<attribute> - (?P<var>{_NAME_RE})(?:(?P<nullish>\?)?\.(?P<member>[^(]+)|\[(?P<member2>[^\]]+)\])\s* + (?P<var>{_NAME_RE})(?: + (?P<nullish>\?)?\.(?P<member>[^(]+)| + \[(?P<member2>{_NESTED_BRACKETS})\] + )\s* + )|(?P<indexing> + (?P<in>{_NAME_RE})(?P<in_idx>\[.+\])$ )|(?P<function> (?P<fname>{_NAME_RE})\((?P<args>.*)\)$ )'''.format(**globals()), expr) @@ -1031,10 +1115,11 @@ class JSInterpreter(object): elif left_val in (None, JS_Undefined): raise self.Exception('Cannot index undefined variable ' + m.group('out'), expr=expr) - indexes = re.split(r'\]\s*\[', m.group('out_idx')) - for i, idx in enumerate(indexes, 1): + indexes = md['out_idx'] + while indexes: + idx, indexes = self._separate_at_paren(indexes) idx = self.interpret_expression(idx, local_vars, allow_recursion) - if i < len(indexes): + if indexes: left_val = self._index(left_val, idx) if isinstance(idx, float): idx = int(idx) @@ -1079,14 +1164,17 @@ class JSInterpreter(object): if md.get('indexing'): val = local_vars[m.group('in')] - for idx in re.split(r'\]\s*\[', m.group('in_idx')): + indexes = m.group('in_idx') + while indexes: + idx, indexes = self._separate_at_paren(indexes) idx = self.interpret_expression(idx, local_vars, allow_recursion) val = self._index(val, idx) return val, should_return - op_result = self.handle_operators(expr, local_vars, allow_recursion) - if op_result: - return op_result[0], should_return + separated = self._separate_at_op(expr) + if separated: + op, separated, right_expr = separated + return self._eval_operator(op, op.join(separated), right_expr, expr, local_vars, allow_recursion), should_return if md.get('attribute'): variable, member, nullish = m.group('var', 'member', 'nullish') @@ -1107,13 +1195,15 @@ class JSInterpreter(object): def eval_method(variable, member): if (variable, member) == ('console', 'debug'): if Debugger.ENABLED: - Debugger.write(self.interpret_expression('[{}]'.format(arg_str), local_vars, allow_recursion)) + Debugger.write(self.interpret_expression('[{0}]'.format(arg_str), local_vars, allow_recursion)) return types = { 'String': compat_str, 'Math': float, 'Array': list, 'Date': self.JS_Date, + 'RegExp': self.JS_RegExp, + # 'Error': self.Exception, # has no std static methods } obj = local_vars.get(variable) if obj in (JS_Undefined, None): @@ -1121,7 +1211,7 @@ class JSInterpreter(object): if obj is JS_Undefined: try: if variable not in self._objects: - self._objects[variable] = self.extract_object(variable) + self._objects[variable] = self.extract_object(variable, local_vars) obj = self._objects[variable] except self.Exception: if not nullish: @@ -1132,7 +1222,7 @@ class JSInterpreter(object): # Member access if arg_str is None: - return self._index(obj, member) + return self._index(obj, member, nullish) # Function call argvals = [ @@ -1275,7 +1365,8 @@ class JSInterpreter(object): assertion(len(argvals) == 2, 'takes exactly two arguments') # TODO: argvals[1] callable, other Py vs JS edge cases if isinstance(argvals[0], self.JS_RegExp): - count = 0 if argvals[0].flags & self.JS_RegExp.RE_FLAGS['g'] else 1 + # access JS member with Py reserved name + count = 0 if self._index(argvals[0], 'global') else 1 assertion(member != 'replaceAll' or count == 0, 'replaceAll must be called with a global RegExp') return argvals[0].sub(argvals[1], obj, count=count) @@ -1316,7 +1407,7 @@ class JSInterpreter(object): for v in self._separate(list_txt): yield self.interpret_expression(v, local_vars, allow_recursion) - def extract_object(self, objname): + def extract_object(self, objname, *global_stack): _FUNC_NAME_RE = r'''(?:{n}|"{n}"|'{n}')'''.format(n=_NAME_RE) obj = {} fields = next(filter(None, ( @@ -1337,7 +1428,8 @@ class JSInterpreter(object): fields): argnames = self.build_arglist(f.group('args')) name = remove_quotes(f.group('key')) - obj[name] = function_with_repr(self.build_function(argnames, f.group('code')), 'F<{0}>'.format(name)) + obj[name] = function_with_repr( + self.build_function(argnames, f.group('code'), *global_stack), 'F<{0}>'.format(name)) return obj diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py index ac1e78002..c4262936e 100644 --- a/youtube_dl/utils.py +++ b/youtube_dl/utils.py @@ -4204,12 +4204,16 @@ def lowercase_escape(s): s) -def escape_rfc3986(s): +def escape_rfc3986(s, safe=None): """Escape non-ASCII characters as suggested by RFC 3986""" if sys.version_info < (3, 0): s = _encode_compat_str(s, 'utf-8') + if safe is not None: + safe = _encode_compat_str(safe, 'utf-8') + if safe is None: + safe = b"%/;:@&=+$,!~*'()?#[]" # ensure unicode: after quoting, it can always be converted - return compat_str(compat_urllib_parse.quote(s, b"%/;:@&=+$,!~*'()?#[]")) + return compat_str(compat_urllib_parse.quote(s, safe)) def escape_url(url): diff --git a/youtube_dl/version.py b/youtube_dl/version.py index b82fbc702..c70d9d2af 100644 --- a/youtube_dl/version.py +++ b/youtube_dl/version.py @@ -1,3 +1,3 @@ from __future__ import unicode_literals -__version__ = '2021.12.17' +__version__ = '2025.04.07' |